Differentiable Multi-Target Causal Bayesian Experimental Design

TL;DR

This paper presents a gradient-based method for Bayesian optimal experimental design to learn causal models efficiently, especially for batch interventions involving multiple targets, outperforming existing heuristics.

Contribution

It introduces a novel end-to-end gradient optimization approach for multi-target interventions in causal discovery, eliminating the need for black-box and greedy methods.

Findings

Outperforms baseline methods in synthetic datasets

Efficiently optimizes over batch multi-target interventions

Applicable to both single-target and multi-target settings

Abstract

We introduce a gradient-based approach for the problem of Bayesian optimal experimental design to learn causal models in a batch setting -- a critical component for causal discovery from finite data where interventions can be costly or risky. Existing methods rely on greedy approximations to construct a batch of experiments while using black-box methods to optimize over a single target-state pair to intervene with. In this work, we completely dispose of the black-box optimization techniques and greedy heuristics and instead propose a conceptually simple end-to-end gradient-based optimization procedure to acquire a set of optimal intervention target-state pairs. Such a procedure enables parameterization of the design space to efficiently optimize over a batch of multi-target-state interventions, a setting which has hitherto not been explored due to its complexity. We demonstrate that our proposed method outperforms baselines and existing acquisition strategies in both single-target and multi-target settings across a number of synthetic datasets.